Vol.2 No.1 2009

74/88

Research paper : Advanced in-silico drug screening to achieve high hit ratio (Y. Fukunishi et al.)−71−Synthesiology - English edition Vol.2 No.1 (2009) as virtual screenings as well as in consultation business. He is a member of the Information Processing Society of Japan and was primarily in charge of developing protein-compound interaction matrices in this study.Kohta SakaiMr. Sakai received his M.S. from the Macromolecular Chemistry Course, Department of Chemistry, Graduate School of Science, Kyushu University, in 1989, and joined Fujitsu Kyushu System Engineering, Limited in the same year. He was primarily in charge of developing 3D structures from the cataloged compounds in this study.Hiroshi KusudoMr. Kusudo joined Hitachi East Japan Solutions, Ltd. in 2002 and is currently engaged in the development of parallel computation systems and research support business. He is a member of the Information Processing Society of Japan and was primarily in charge of developing protein-compound interaction matrices in this study.Haruki NakamuraDr. Nakamura received his Ph.D. from the Graduate School of Science, the University of Tokyo, in 1980. He was an assistant professor at the Faculty of Engineering, the University of Tokyo, and served at the Protein Engineering Research Institute and the Biomolecular Engineering Research Institute. He has held the position of professor at the Institute for Protein Research, Osaka University, since 1999. He specializes in biophysics and was primarily in charge of collecting public data sets and supervising the overall research in this study.Discussion with Reviewers1 Significance of developing the compound databaseQuestion and comment (Akira Ono)The authors stated a clear research aim and comprehensibly portrayed a scenario to select elemental technologies as shown in Fig. 1, which were then integrated into a practically operational database. This is a typical Type 2 Basic Research and is also an excellent example of Product Realization Research. It is expected that the database developed in this study will be highly valued through its applications to drug screening for target proteins.Answer (Yoshifumi Fukunishi)A compound database is a repository of knowledge regarding the “synthetic easiness” of compounds synthesized in the past. We expect that the database will not only be directly used for drug discovery but also function as a basis to understand easily synthesizable compounds and even previously inconceivable compounds, which may lead to establish a new research field.2 Design of unknown active compoundsQuestion and comment (Akira Ono)The compound database is intended for efficiently examining the strength of chemical bonds between a particular target protein and numerous predefined compounds, thereby dramatically improved a hit ratio of drugs. Now, is it possible for the users to predict unknown compounds that could bind with a particular target protein even more strongly by using the database?Answer (Yoshifumi Fukunishi)There is a possibility that the users can predict previously unknown active compounds that are not included in the compound database. There was a trend that active compounds were classified into several clusters according to their chemical properties as a result of screening compounds that potentially bind to a particular target protein. Therefore, it is considered feasible to synthetically design an unknown active compound that exhibits the properties of a particular active compound group.3 Possible improvement of the databaseQuestion and comment (Akira Ono)As mentioned in section 4.3, it is important to recognize that excessive pursuit of mathematical strictness is meaningless in developing a compound database. In this regard, is there still a room for optimizing the database according to needs by reassessing its development process?Answer (Yoshifumi Fukunishi)Redesigning of the database according to needs, i.e., ad-hoc database, is considered possible. For example, we adopted a dominant configuration of compounds in water (regarding the protonation state of a carboxylic acid, for example, -COO- was adopted instead of -COOH) for the current version of the database; however, the molecular configuration could be changed when the compounds bind with proteins. Recently, it is often discussed that a docking simulation of a compound targeting a highly-charged protein pocket is extremely difficult. A dominant ion form of carboxylic acid could occasionally be -COOH inside the negatively-charged protein pocket. Hence, it will be important to develop a target-oriented compound database in future.4 Comparison with existing overseas databasesQuestion and comment (Akira Ono)It is mentioned in section 4.11 that when the current database was employed, the enrichment factor was improved approximately 40- or 70-fold compared with random screening. In contrast, how superior is the developed database over the precedent overseas compound databases in terms of enrichment or enrichment factors?Answer (Yoshifumi Fukunishi)In general, unsuccessful predictions of computationally screened compounds are not published as research articles; thus, it is difficult to compare databases in detail. In our case, the hit ratio was 3–30 % when computationally predicted 100–300 compounds were purchased. So far, only one out of five cases showed 0 % hit ratio. Hit ratios reported on other literatures are mostly 10 % at best and 50 % of targets show 0 % hit ratio. Therefore, the developed database combined with our prediction method is considered more effective than the existing overseas databases.5 Consideration of tautomers and ion formsQuestion and comment (Takatsugu Hirokawa)While commercially-produced overseas databases and software are acquiring the major share in the field, it is noteworthy that such a high-quality compound database and an unprecedentedly unique protein-compound affinity matrix were released by Japanese researchers. Various issues involved in digitally processing compound data were fully addressed in each development process discussed in the article, and this ensures that the database can be used reliably by researchers. This work should also be highly acclaimed as Product Realization Research. Regarding the protonation of compounds, how are the tautomers and ion forms considered (such as whether pH7.0 is assumed) besides the statement, “We investigated the protonatation states of various functional groups...under a vacuum and in water (near pH7.0)” of section 4.4.Answer (Yoshifumi Fukunishi)Protonation status is based on the assumption of pH7.0. However, since it is difficult to predict accurate pKa, a dominant configuration of each functional group contained in a molecule at pH7.0, rather than the pKa of the whole molecule, was adopted.

※このページを正しく表示するにはFlashPlayer9以上が必要です